Removal of baseline noise from Electrocardiography (ECG) signal based on time domain approach

Electrocardiographic (ECG) signals in measurements are often contaminated with different types of noises in which include baseline noise. In case of the frequency of baseline noise is greater or smaller than frequency of the ECG signal, it is easy to filter the baseline noise from ECG signal by using filtered methods in frequency domain. In contrast, if frequency of baseline noise and frequency of ECG signal are coincident, it is difficult to apply the frequency domain filters for baseline noise removal. In this paper, we introduce an approach to remove of baseline noise from ECG signal in time domain and evaluation the efficacy of the method based on Mean Square Error criteria. We have performed experiments with simulated ECG signal which including white noise, random and sinusoidal baseline noises. Throughout the experiment, we found that the errors of time domain filters depend on the amplitude of the base line noises.